Author: Chessfun
Date: 09:04:58 09/30/02
Go up one level in this thread
On September 30, 2002 at 08:08:09, Dadi Jonsson wrote: >On September 29, 2002 at 20:20:37, Chessfun wrote: > >>On September 29, 2002 at 20:09:55, Dadi Jonsson wrote: >> >>>I installed Ruffian 1.0.1 under the Chess Assistant interface (UCI) and ran two >>>matches between Ruffian and Gambit Tiger 2 (using CA's Engine-Engine). The only >>>difference between the two matches was a different opening book for Ruffian. >>> >>>AMD 1.4GHz >>>512MB memory >>>128MB hash >>>pondering off >>>Tablebases: 3, 4 and some 5-man. >>>1 minute + 1 sec. per move. >>> >>>=== Match 1: Ruffian uses its own book >>> >>>It was a close match with Ruffian leading after 40 games, but Gambit Tiger >>>winning in the end. I watched some of the games and noticed that 10-15 moves >>>into the game Ruffian was usually about 10 seconds behind on the clock. >>> >>>Tiger Gambit II: 80, +30=23-27 (41.5/80) 51% >>>Ruffian : 80, +27=23-30 (38.5/80) 48% >>> >>>For comparison here are the results of Chessfun's match between the same engines >>>(same time-limit) running under the ChessBase interface: >>> >>>Ruffian - Gambit Tiger 2.0 : 40 (+ 12,= 10,- 18), 42.5 % >> >> >>In my case the game was played also using Tiger.ctg from Chessbase. >>And as you say at the end in this case you used the CA book against Ruffians >>book. >> >>Sarah. >As you said in your message on September 28 "All engines using their own books >excluding, Crafty, Goliath using general.ctg". So you used Tiger.ctg for the >Tiger and presumable Ruffian's own book. Reason for stating the book was it's possible to use Ct.tbk in chessbase simply by creating an empty.ctg book and putting Ct.tbk in the engine directory. And naturally Ruffian used it's own book. >It's important to realize that although >it doesn't affect the conlusion of the experiment. >My point with the experiment was that Ruffian is doing amazingly well *in spite* >of being handicapped by its opening book. By giving it a better opening book we >can expect to see a jump in its performance. I don't actually see how you can come to the conclusion with a few lightning games. I have played some 150 tournament games with Ruffian and after looking at all those games, IMO It's book is very strong. I'm sure I could play another 40 lightning games with the exact same conditions and get a different result as both books have more than 40 lines in them. Currently I have it slightly below CM 9000 in strength 4 points and for a free program it is an amazing performer. Although I don't see even giving it Fritz 7.ctg would change it's performance by more than +20 points. Also the same could be said for a lot of programs. (12) Ruffian 1.0.0 : 150 (+ 51,= 49,- 50), 50.3 % Hiarcs 7.32 : 10 (+ 4,= 3,- 3), 55.0 % Chess Tiger 14.0 : 10 (+ 3,= 2,- 5), 40.0 % Gambit Tiger 2.0 : 10 (+ 4,= 2,- 4), 50.0 % Goliath Light 1.5 : 10 (+ 8,= 1,- 1), 85.0 % Fritz 7 : 10 (+ 6,= 1,- 3), 65.0 % Gandalf 5.1 : 10 (+ 3,= 4,- 3), 50.0 % Shredder 6 : 10 (+ 3,= 3,- 4), 45.0 % Tiger 15 normal style : 10 (+ 1,= 4,- 5), 30.0 % Hiarcs 8 : 10 (+ 5,= 3,- 2), 65.0 % Shredder Paderborn : 10 (+ 2,= 5,- 3), 45.0 % Crafty 18.15 : 10 (+ 4,= 5,- 1), 65.0 % Deep Fritz 7 : 10 (+ 0,= 5,- 5), 25.0 % Chessmaster 9000 : 20 (+ 6,= 8,- 6), 50.0 % Tiger 15 gambit agressive : 10 (+ 2,= 3,- 5), 35.0 % Sarah.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.